The major objective is to develop a methodology of decision making for patients with acute leukemia. The aim is to utilize all relevant clinical and laboratory data during the time course of the disease to predict therapeutic and toxic results of various schedules and doses of treatment. Regimens of treatment will be suggested to the physicians based upon continuing analyses of data from a given patient and from all patients receiving the same treatment. A working system has been developed for the collection, storage, retrieval and analysis of clinical and laboratory data from patients with acute leukemia. The system will be modified to utilize the computer more efficiently and to permit input of data from multiple sources, such as from a medical record analyst, a research nurse, the laboratory of clinical pathology, laboratories doing special studies (infection studies, immunology studies) and other sources. The continuing analyses will include: characterizations of the time course of disease for individual patients; analyses and summaries of data on a regular basis for a group of patients on one regimen of treatment; and long term studies to determine characteristics of patients related to prognosis for patients with acute leukemia. The entire system will be designed for use in multiple hospitals and clinics so that comparable studies and analyses can be performed. It is proposed to develop new methodology for the analysis of data from patients with acute leukemia based on techniques of multivariant analysis and multivariant time series.